Monitoring and Assessing Arterial Traffic Performance

Transcription

1 Monitoring and Assessing Arterial Traffic Performance Stanley E. Young, P.E. Ph.D. University of Maryland Center for Advanced Transportation Technology Traffax Inc. National Renewable Energy Laboratory

2 Outline Traffic Data Fidelity of Outsource Probe Data Where we have been, where we are now. Completing the Picture Arterial Performance Measures Bringing in Volume Data State Wide Extending Real-Time to Arterial Networks Its time for Arterial Management Systems

3 I-95 Vehicle Probe Project Phase I ( ) First Probe-based Traffic System Specifications-based, validated Licensing - one buys, all share Began 2.5K miles, grew to 40K Travel time on signs, 511 systems, operational awareness, performance measures Phase II (2014 forward) All of the above Better quality, less cost Data market place (Multiplevendors) Emphasis on arterials and latency 42.5K and growing Map-21 Performance Measures

4 Vehicle Probe Project Phase I ( ) First Probe-based Traffic System Specifications-based, validated Licensing - one buys, all share Began 2.5K miles, grew to 40K Travel time on signs, 511 systems, operational awareness, performance measures Phase II (2014 forward) All of the above Better quality, less cost Data market place (Multiplevendors) Emphasis on arterials and latency 42.5K and growing Map-21 Performance Measures

5 First Multi-Vendor Freeway Validation I-83 & I-81 Harrisburg, Oct 2014 PA Segments 31.3 miles Data collection 2300 to 2555 total hrs 71 to 80 hrs [0-30] 53 to 66 hrs [30-45] AASE 2.1 to 4.1 mph [0-30] 3.1 to 5.8 mph [30-45] 5 May 07, 2015

6 PM Peak Hour (Oct 15-16, 2014) 6 May 07, 2015

7 Non-recurring Congestion Oct 13, AM to 7 PM 7 May 07, 2015

8 Arterial Probe Data Quality Study 2013 mid 2014 State / Set ID Road Number Road Name Validation Date Span # of Segments # of Through Lanes AADT Range (in 1000s) Length* (mile) # Signals / Density # of Access Points Median Barrier Speed Limit (mph) NJ-11 NJ-12 PA-05 PA-06 VA-07 VA-08 MD-08 US-1 Trenton Fwy, Brunswick Pike Sep 10-24, / Yes NJ-42 Black Horse Pike / Yes US-130 Burlington Pike / Yes 50 NJ-38 Kaighn Ave / Yes 50 Nov 5-19, Palmyra Bridge NJ / Yes Rd. US-1 Lincoln Highway / Yes Dec 3-14, Conchester US / No Highway PA-611 Easton Rd / NO PA-611 Old York Rd Jan 9-22, / Partial PA-611 N Broad St / NO VA-7 US-29 US-29 MD-140 Leesburg Pike and Harry Byrd Hwy Lee Hwy (S Washington St) April 5-16, Case Studies from Spans NJ through NC Test extent of probe data 15K AADT to 100K 2 12 lanes / Yes / Partial 30 Lee Hwy May 8-19, / Partial April 30, Reistertown Rd June 5-14, to signals / 3.8 per 148 mile No Baltimore Blvd / YES Objective: Reference case studies 8

9 Arterial Probe Data Recommendations Likely to have usable probe data <= 1 signals per mile AADT > Fully or Partially captures >75% slowdowns Possibly usable probe data <= 2 signals per mile AADT 20K to 40K May Fail to capture > 25% of slowdowns Should be tested Likely not usable probe data >=2 signals per mile Not recommended 2013/14 Data not ready for Prime Time Probe data quality most correlated to signal density Consistent positive bias at low speeds As probe data improves, delay will increase Other challenges include: Severe queuing, multi-cycle failures, 9 Optimistic bias in bi-modal traffic Insensitive to signal timing changes April 30, 2015

10 Roadmap for Arterial Management Systems Arterial Performance Measures are fundamentally different than Freeway Performance Measures Until recently (2014), performance assessment has been too costly for broad based monitoring and performance measures New technology developments have enabled first generation large scale performance assessment Include Re-identification data, High-Resolution Controller data We are NOW (2015) with arterials, where we were in 2008/9 with freeways

11 Technologies Enabling Arterial Management Systems Re-identification High-Res Signal Data Both enabled by consumer wireless communication and big data processing. Available Now Multiple Vendors - Cost Effective Direct samples vehicle travel time (5% for BT) Logs all actuation and phasing information Works best at corridor level Works at intersection level Independent of Signal System Provides top-level user experience information Integrated with Signal System Provides detailed intersection analysis and data for optimizing signal system Not one or the other but both!

12 Emerging Arterial Performance Measures Travel Time and Travel Time Reliability based on sampled travel time sources Enabled by re-identification data, later outsourced probe data and connected vehicle data as it matures Fundamentally linked to the statistical distribution of travel time Percent Arrivals on Green reflects quality progression Supported by methods such as Purdue Coordination Diagram tools Split Failures (frequency of occurrences) Reflects capacity constraints Related to GOR / ROR

13 Re-Identification Data (Bluetooth) Uses a ID unique to a vehicle (MAC ID of a Bluetooth device inside vehicle) An initial detector identifies when a vehicle enters a corridor by the vehicle s ID Another detector reidentifies the vehicle at the end of the corridor Travel time/ speed can be directly calculated from the entry and exit time Direct samples of Travel Time Picture source: libelium.com

14 Travel Time and Travel Time Reliability Based on directly sampled travel time measurements For arterials, can be applied. Intersection to intersection Corridor based Network level, origin to destination Directly reflects concerns of the traveling public Efficient and predictable travel Measures can be applicable to other modes of travel Freeway, transit, air, etc.

15 Re-id Travel Time Data Fidelity

16 CFD Statistical Performance Measures

17 CFDs to Contrast Performance 100% 90% 80% 70% 60% 50% Comparative CFD 40% 30% 20% Before After 10% 0% Travel-time Minutes

18 CFDs to Contrast Performance 100% 90% 80% 70% 60% 50% Comparative CFD 40% 30% 20% Before After 10% 0% Travel-time Minutes

19 Percent Arrival on Green and Split Failures Percent Arrivals on Green Measure on how effectively signals are coordinated, moving vehicles through the system The higher the PAG, Less stops, happier customers Higher corridor speed, better fuel economy, less emissions Direct indicator of signal system performance Split Failures (i.e. Capacity Constraint) Measures percent of system (time and space) suffering from lack of capacity The need more capacity metric, or get off my back metric, its time to spread the pain metric... Something more than signal optimization required capacity/demands need to be addressed

20 High Resolution Signal Data Logging of sensor and phase information Data forwarded periodically to central server Applications Purdue Coordination Diagram Red-Occupancy Ration / Green Occupancy Ratio Volume / Demand Analysis (per movement) Streamlined Maintenance Picture Source: FHWA

21 High Resolution Signal Data Logging of sensor and phase information Data forwarded periodically to central server Applications Purdue Coordination Diagram Red-Occupancy Ration / Green Occupancy Ratio Volume / Demand Analysis (per movement) Streamlined Maintenance Picture Source: FHWA THIS IS CONNECTED INFRASTRUCTURE!!!!!

22 Sample Metric - PAGs Purdue Coordination Diagram Time In Cycle (s) Time In Cycle (s) :00 16:00 17:00 18:00 Time of Day 0 15:00 16:00 17:00 18:00 Time of Day

23 PAG in the news!

24 Sample Metric - Intersection Movement Capacity Analysis (ROR GOR) 1 1 Red Occupancy Ratio Red Occupancy Ratio Green Occupancy Ratio Green Occupancy Ratio

25 Frequency of Split Failures Indicator of oversaturation When demand overruns capacity Indicates when additional capacity or demand management is required Also known as the metric for. Get off my back, nothing left to do Time to share the pain Give me another lane if you want this solved

26 Current State of Arterial Management Systems (AMS) Integration of Management Systems And Decision Making Integration of Management Systems And Engineering Practices Pavement Management Systems Development of Formal Data-Driven Management Systems Standard Data Collection Methodologies Developed Arterial Management Systems Consensus Established on Performance Measures Exploratory Research on Performance Measures 1950s 1960s 1970s 1980s 1990s 2000s 2010s 2020s

27 Challenges / Benefits to Arterial Performance Measures Created a common lexicon/language Between Traffic, Ops, Planning Define Performance Levels (Good, Mediocre, and Ugly) Effective communication with management and public Systematic approach Link performance to budget/funding Long term performance tracking Predictable return on investment Linking to other Priorities Operations during freeway incidents Energy efficiency, dglobal warming (GHG emissions)

28 Real-Time Arterial Performance Stuff from Chapter 1

29 Conclusions Final Thoughts Arterial Performance Fundamentally Different than Freeways Re-identification and Hi-Res Data enable full observability Key Measures Include Travel time and travel-time reliability Quality of progression Degree of capacity saturation These Enable Performance Management of Arterials

30 Thank You! Stanley E. Young, P.E. Ph.D. University of Maryland Center for Advanced Transportation Technology Traffax Inc. National Renewable Energy Laboratory Mobile